Skyscanner Travel Planner Snafu Highlights the Need for Live Agent–Bot Balance



In the age of Siri, Cortana, and Echo, the idea of artificial intelligence conjures up a sense of fear not because of how it's been portrayed in science fiction, but because AI may one day pose a very real job threat across a number of professions, including customer service. AI tools can perform certain tasks faster and more efficiently than people, but a recent customer service hiccup from travel planning company Skyscanner serves as a reminder that, however powerful it is, AI won’t be a catch-all tool for customer support.

When customer James Lloyd got a suggested itinerary from Skyscanner’s automated travel planner that included a 47-year layover in Bangkok, he sent the company a message on Facebook, humorously asking what he should do with all his time in Bangkok. Human customer service representative Jen then responded in a way that most modern artificially intelligent service tools can’t—with wit. Jen listed a number of things Lloyd could do in Bangkok for 47 years, including taking a cruise on the Chao Phraya river and becoming a Tai Chi expert in Lumpini Park, but included a note at the bottom, promising to have someone look into the layover error. She not only solved the problem, but recognized the irony of it.

“The climax of the classic movie 2001: A Space Odyssey came when the main computer, Hal, took control of the spaceship with sinister intent. The astronaut regains control, not by outsmarting Hal, but by outwitting Hal. [Artificial intelligence] is like Hal, long on smarts but short on ingenuity and intuitive reasoning. We will always need human reps with their unique capacity to be creative to remain in control,” says Chip Bell, customer service expert and author.

The key part of the Skyscanner story is the capacity of service rep Jen to intuitively read the playful tone of the customer’s query and provide a clever, funny response that yielded social media accolades, according to Bell. When service recovery is laced with empathy and compassion—features unavailable to computers—it provokes a heart-tugging story likely to go viral, he explains. “Research shows customers with a problem or issue that is remarkably solved end up more loyal than customers who have never even had a problem.  Before great recovery, customers operate on hope and faith—after great recovery, they have enhanced confidence through concrete proof,” he adds.

Social media is a conversation, says Natalie Petouhoff, vice president and principal analyst at Constellation Research, and having contextual relevancy is key to making that conversation relatable, maybe even funny. "That’s why having a human answer a customer question can be far more satisfying than a bot. It's interesting the bot made a mistake of 47 years—that’s an oops—but a live agent took that oops and turned it around. Perhaps someday, maybe not in my lifetime, bots will have that contextual relevancy to either not suggest 47 years layover or see the humor in it and become able to respond like the live agent did. But it's not likely anytime soon," she says. 

Related Articles

Helpshift's cofounder and CEO, Abinash Tripathy, maintains that omnichannel service is not the answer to customers' growing expectations. As customers increasingly expect faster, more efficient support, the need for more sophisticated self-service support and support through artificial intelligence will grow. With $23 million in funding, Helpshift plans to continue improving its in-app messaging customer service tool by making it more proactive.

Posted June 13, 2016

Inbenta's Chatbot Development Platform enables businesses to deploy artificial intelligence customer support for Web sites, Facebook Messenger, and Skype.

Posted July 13, 2016

The virtual assistant will provide visitors to IP Australia's Web site with information about the trademarking process.

Posted August 11, 2016

DigitalGenius' Human+AI blends human and artificial intelligence for the contact center.

Posted August 11, 2016